abstract. The reform of contributory negligence into a scheme of apportioning liability based on comparative fault is among the most significant developments in tort law during the Twentieth Century. Its significance goes beyond the rejection of the common law's “all or nothing” attitude about liability and has extended to modification of many other aspects of tort law that developed because of the entrenchment of contributory negligence.

At about the same time as courts and legislatures adopted comparative fault, the advent of large toxic substances case congregations emerged – asbestos, Agent Orange, DES, silicone gel breast implants, and tobacco are among the most notable of such. Many of these cases present difficult problems of causation because the connection between exposure to the agent and disease is only dimly understood. The best scientific evidence is provided by epidemiology, which is group-based and statistical in nature.

The confluence of comparative fault principles and probabilistic evidence of causation raises the question of whether liability should be imposed proportionally based on the probability of causation in toxic substances cases. Many scholars have argued for rejection of the customary “more likely than not” standard for the burden of proof and for adoption of a proportional liability rule. This chapter critically assesses those proposals by looking carefully at the precision and fallibility of the epidemiological evidence on which the scholars' proposals rely. After concluding that proportional liability would not provide the deterrence benefits claimed for it, this chapter considers the implications of its analysis for employing proportional liability in other areas of tort law.